ArticlePDF Available

Managerial decision support tools-different perspective

Authors:

Abstract and Figures

Manager who is adjusted to think about decision process in the terms of intuitive process, rational model, and model of bounded rationality is usually confused with many technical details describing something he does not understand. To take more advantage of decision support tools managers need IT professionals to speak with the same language. This article represents an attempt how to use managerial language in order to describe decision support tools.
Content may be subject to copyright.
91
MANAGERIAL DECISION SUPPORT TOOLS – DIFFERENT
PERSPECTIVE
Hana Kopáčková
Ústav systémového inženýrství a informatiky, FES, Univerzita Pardubice
Abstract
Manager who is adjusted to think about decision process in the terms of intuitive process,
rational model, and model of bounded rationality is usually confused with many technical details
describing something he does not understand. To take more advantage of decision support tools
managers need IT professionals to speak with the same language. This article represents an
attempt how to use managerial language in order to describe decision support tools.
Keywords
Decision, decision support tools, gathering information, model of bounded rationality, rational
model.
1. Introduction
Decision-making process represents undoubtedly an essential and indispensable part of human
life. Every day brings new problems that must be solved, accompanied by decisions having
different impact. Some these problems need fast solution whence act of decision-making is done
by individual, mostly using intuition and previous experience.
o
Intuitive decision-making, according to Thagard [12], has three main advantages:
o
Speed – an emotional reaction can be immediate and lead directly to a decision.
o
Interest basing decision on emotions helps to ensure that the decisions take into account
what decision maker really care about.
o
Action – the positive feeling toward an option will lead directly to action.
Nevertheless emotion based intuitive decision-making can also have some serious
disadvantages. An option may seem emotionally appealing because of failure to consider other
available options. Another problem with intuition is that it may be based on inaccurate or
irrelevant information. Finally, intuitive reasoning is problematic in group situations where
decisions need to be made collectively.
Intuitive decision-making is purely based on real person who acts as decision-maker. Due to
this fact no ICT tools can be used to make such process easier and it is also reason why I will not
cover this branch in more detail. Moreover managerial decision-making should not be so
dependent on one person.
Theory of decision-making has taken, as its subject matters how individuals and groups make
decisions [2], [4], [5]. The goal for much of this work has been the production of a model of
decision making - a model general enough to describe individual cases of decision making while
drawing out important generalities across different individuals and situations.
One example of decision-making model, that can be described and supported by information
and communication technologies, represents rational model. It assume that a decision maker
92
posses a utility function (an ordering by preference among all the possible outcomes of choice),
that all the alternatives among which choice could be made were known, and that the
consequences of choosing each alternative could be ascertained [9], [11].
Unlike intuitive decision-making, this type of decision-making process can be described and
divided into steps, for example according to Bazerman [1]:
o
Define the problem, characterizing the general purpose of your decision.
o
Identify the criteria, specifying the goals or objectives that you want to be able to
accomplish.
o
Weight the criteria, deciding the relative importance of the goals.
o
Generate alternatives, identifying possible courses of action that might accomplish your
various goals.
o
Rate each alternative on each criterion, assessing the extent to which each action would
accomplish each goal.
o
Compute the optimal decision, evaluating each alternative by multiplying the expected
effectiveness of each alternative with respect to a criterion times the weight of the criterion,
then adding up the expected value of the alternative with respect to all criteria.
Although the assumptions of rational model (results and alternatives are ascertained,
information are complete and so on) cannot be satisfied even remotely for most complex
situations in the real world, they may be satisfied approximately in some isolated problem
situations without uncertainty.
The computational tool of linear programming, which is a powerful method for maximizing
goal achievement or minimizing costs while satisfying all kinds of side conditions, can find
optimal result within the limits of approximation of such model to real world conditions.
Problem of scheduling, time tabling and routing can be also covered in the rational model of
decision-making due to certainty of results. Methods used to solve these problems range from
pure statistical methods to artificial intelligence methods (genetic algorithms, neural networks...).
All these methods are now applied by computers reducing complexity and time requirements.
Although examples of rational decision-making process are not rare, limits of incomplete
information, inconsistency, and institutional constraints on alternatives brings us to the models of
limited (bounded) rationality. These models retain the same process of decision-making but
incorporate many additional limitations. Simon [10] describes ‘satisfying’ behavior - a decision
making process which searches for ‘good enough’ options, rather than an optimum solution. With
satisfying decision making becomes something which is carried out in a limited time, and with
some limits on the individuals concerned.
Theories of bounded rationality have been developed to incorporate the importance of rules
and identities in decision making. Rules and identities are important logics which are used in
deciding what to do - the so called ‘institutional’ aspects of activity.
93
Figure 1: Steps of decision making in bounded reality. Freely adapted from [3]
2. Problem formulation
Model of bounded reality covers the most frequent decision making situations in the real
world. Taking this model as a basic framework, now I want to introduce great amount of tools
from the range of information and communication tools that can be useful in decision making
process.
Fig. 1 gives a pictorial summary of necessary steps of the individual and group decision
making. It also depicts the inclusion and integration of personal and institutional aspects of such
activity. The sections that follow describe each step of individually looking for possible ICT
tools.
Identification of the problem to be solved can be very simple in the case of allocation of tasks.
If the problem is allocated to manager’s responsibility, he/she can continue with the step two
without usage of any tool. But the situation doesn’t have to be so clear. Managers are supposed to
be proactive and look for sources of possible future dangers. In this situation ICT tools can be
used in their diversity, starting from text documents, spreadsheets, databases to methods of
Understanding of the problem
Setting managerial objectives
Individual or group
decision making?
Group
Individual
decision
Identification of the problem to be solved
Choice of appropriate
decision making style
Allocation of tasks
Group
Communication
Comparison and evaluation of alternatives
Searching for consequences
The act of
choice
Creation of alternatives
Implementation of decision
Setting criteria for evaluation of
alternatives
NO – change managerial objectives YES
Follow-up and control Results are satisfying?
Information,
personal attitudes,
organisational
rules
94
machine learning. Especially methods of unsupervised learning can be used to find something
new. Unsupervised learning is a method of machine learning where a model is fit to
observations. It is distinguished from supervised learning by the fact that there is no apriori
output. Such techniques are adopted in data mining tools enabling even managers without
previous IT education to use them. This is very important presumption for all ICT tools that can
be used in decision making. Working with these tools must be clear and user friendly otherwise
they remain as the nice technical gadgets staying on the shelf.
Understanding of the problem is composed of detailed description of the problem, separation
of unique features of the problem, specification of changes and testing of causes. This step of
decision making is very important. Even small mistake in setting of causes can overthrow the
whole process of decision making. Results that would be found would not lead to satisfying
solution of the problem and the whole process would have to necessary start again. Due to these
facts, managers should not pass this step without usage of any tools. For example modeling tools
normally employed for business process modeling can be used in this step if companies have
them. If it is not possible, any drawing program can serve in the similar way. In the case of very
important and sensitive problem, expert systems provide managers with best information. An
expert system is regarded as the embodiment within a computer of a knowledge-based
component from an expert skill in such a form that the system can offer intelligent advice or take
an intelligent decision about a processing function. A desirable additional characteristic, which
many would consider fundamental, is the capability of the system, on demand, to justify its own
line of reasoning in a manner directly intelligible to the enquirer [13]. Main problems with expert
system rise from their advantage. Due to their complexity, creation of the expert system is very
lengthy and demanding which is adequate to its price. Application of such tool is therefore
evincible only in the case where claims committed by wrong decision are too high.
Setting of managerial objectives is one of the steps which are predisposed to personal attitudes
and organizational rules. Rational model of decision making assume maximizing of utility
function but organization may prefer alternative objectives or manager as a decision maker may
prefer different strategy according to previous practice. Influence of these institutional aspects is
in the most cases necessary and it does not have to have negative impact. Unfortunately in this
step managers can not use any ICT tool to support decision making.
Afterwards setting managerial objectives, decision maker has to select one from two ways.
Either he can choose to continue in decision making process alone or involve other persons into
this process. Both alternatives have their pros and cons. Great advantage of individual decision
making is time point of view. Individual decisions take less time but on the other hand amount of
information used is determined only by manager’s knowledge. Another disadvantage can be seen
in the influence of personal attitudes that can excessively narrow process of decision making (too
few alternatives, simplified comparison and evaluation of alternatives...). Group decision is on
the other hand much more time-demanding, requiring communication in group. Nevertheless ICT
tools can be very helpful in supporting group communication (instant messaging, video
conference, sms communication, e-mail, and other tools can make communication faster and
easier).
Decision about individual or group solution of problems should not be supported by ICT tools
since the amount of decisive facts is too huge and not stable.
In the event of choosing group decision making, another step represent selection of
95
appropriate decision making style. According to [14] manager can choose one out of five styles.
The way of selection can be represented as a decision tree but with some restrictions (model does
not give explicit recommendation, differentiation among styles is insufficient, and so on). New
version of their model can be described mathematically by seven equations, giving great space to
ICT usage. Manual calculation of such model is very demanding in comparison with computer
application that takes only minimum time and computational capacity.
Allocation of tasks is purely dependant on decision makers, how they distribute tasks, roles
and responsibility. I suggest no usage of ICT in this step.
Following five steps are common for both; group and individual decision. Differences can be
found in the way of seeking solution, gathering information and so on.
Setting criteria for evaluation of alternatives represent such activity where only criteria with
defined qualities are selected. Those requirements are: completeness, lucidity, null redundancy,
and minimal range. Seeking for appropriate criteria should always start with managerial
objectives, considering the fact that these criteria will serve as an indicator of fulfillment of
managerial objectives. For each partial objective should be set minimally one criterion,
nevertheless criteria set in this way are insufficient and must be completed by other ones.
Identification of subjects whose interest can be touched shows us other important criteria which
must be applied in order to make decision acceptable; compatibility with personal interests.
Second thing that must be accepted during the process of setting criteria is possibility of negative
impact of final decision. Such impact is usually not covered in managerial objectives but it is
necessary to formulate it in this stage. From the point of view of ICT tools, we can find only
databases with possible criteria prepared during previous decision making to be used.
Next step in decision making is creation of alternatives. It is a case of demanding creative
process covering systematic and analytical methods and methods giving much space to creative
inspiration. Examples of particular methods are as following: brainstorming, brainwriting, Delft
method, bionics, morphological analysis, decision trees, matrix of hypothesis and so forth.
Except from brainstorming, other methods can be applied by ICT preparing the environment for
straightforward interpretation. If the problem is based on quantitative data (value of sales, number
of claims, results of statistical measurements...) it is possible to engage management information
system (MIS) as indirect tool using preprocessed data or data mining tools that can discover
hidden dependences. Due to this new information, manager is able to change original alternative
or to create new one. The last but not least is possibility of using expert systems, which is the
most expensive but in some cases indispensable. Main problem in this step represent the fact that
all the process is highly dependent upon institutional constraints. The situation is affected by the
same elements as it was described in the step of setting managerial objectives. For the most part
of situations, rational model is then broken. Nevertheless number and quality of alternatives is
very important for the whole decision process.
Searching for consequences is closely connected with creation of alternatives and sometimes it
is mixed together. Process of searching depends on the type of solved problem but in all cases it
must be done according to required criteria. Quantitative problems can be easily solved using
mathematical modeling. In this case we can find many ICT tools which can be used directly for
finding consequences, either these tools are special, only for selected problems, or general. Ill
structured qualitative problems are more dependent on decision maker or other expert.
96
Comparison of alternatives and the act of choice is subject to influence of institutional
constraints.
Following steps are not part of decision making process but I mentioned them here just to see
that even good decision that is poorly implemented can spoil the whole process.
3. Gathering of information
Rational model and model of bounded rationality define steps of decision process without
emphasis on data gathering. It is given as an activity that will be done somehow. I would like to
stress necessity of textual data and to show which methods appears to be effective.
Automated processing of text documents can prevent simplifications and generalisation, which
allow us to decide on the base of small amount of cases and widen this decision on all cases.
Unfortunately, this approach is commonly used having too much text documents and only little
time to read them. Usage of text categorization methods can also significantly lower manager
workload.
Methods suitable to support managerial decision-making process must fulfil these criteria:
o
easy to implement,
o
fast to process categorization,
o
cheap,
o
stable for differences in length of document,
o
learning model build from small number of documents,
o
high precision.
Three methods which are easy to apply were tested for being used in managerial decision-
making. These methods are K-nearest neighbour, Rocchio algorithm and Naive Bayes algorithm.
In the testing environment I used 50 documents in Czech language; 25 of them were focused
directly on the branch of waste management and the rest 25 documents were not specialised
only covered problem of the environment. The shortest document had only 98 words and the
longest had 1400 words. For feature selection were used three different methods: Chi-square,
Mutual information, and Information gain [7], [6] for the weighting of words was used TFIDF
weighting [8]. After the pre-processing stage database was filled with 10571 words.
The result can be seen in Table 1 (CCI means correctly classified instances and K means
Kappa statistics).
Table 1: Effectiveness of classificators
Naive Bayes K-NN Rocchio algorithm
CCI [%]
K [%]
CCI [%]
K [%]
CCI [%]
K [%]
Chi-
square
96,00 92,00
66,00
32,00
100,00
100,00
Mutual information 96,00 92,00
68,00
36,00
100,00
100,00
Information gain 92,00 84,00
50,00
0,00
100,00
100,00
K nearest neighbour algorithm is very sensitive to length differences (documents using same
words have long Euclidean distance between them if they differ in length) so it is not suitable to
support managerial decision-making as it was described here. On the other hand Rocchio
algorithm and Naive Bayes can serve as very helpful tool.
97
4. Acknowledgment
This research and paper was created with a kind support of the Grant Agency of the Czech
Republic, grant number GACR 402/05/P155.
5. References
[1] Bazerman M. H. Judgment in managerial decision making. New York: John Wiley, 1994.
[2] Eppen G., Gould F. Introductory Management Science. New Jersey: Prentice-Hall, p. 164-165, 1984.
[3] Fotr J., Dědina J., Hrůzová H. Manažerské rozhodování. Praha: Ekopress, 2003
[4] Golub A. L. Decision Analysis: An Integrated Approach. New York: John Wiley & Sons, 1997.
[5] Leiw A., Sundaram D. Complex Decision Making Process: Their Modelling and Support. In: Proceedings of the
38th Hawaii International Conference on System Sciences, IEEE, 2005.
[6] Li Y. H., Jain A. K. Clasification of text documents. In: The Computer Journal, vol. 41, no. 8, 1998.
[7] Quinlan J. R. Induction of Decision Trees. In: Machine Learning, vol. 1, no.1, 1986.
[8] Salton G. Developments in Automatic Text Retrieval, In: Science, Vol. 253, 1991.
[9] Simon H. A. A Behavioral Model of Rational Choice. Cowle: Foundation Paper 98. In: The Quarterly Journal of
Economics, vol. LXIX, February 1955.
[10] Simon H. A. Models of Bounded Rationality., Cambridge, M.A.: Harper and Row, 1983.
[11] Simon H. A. Report of the Research Briefing Panel on Decision Making and Problem Solving. National
Academy of Sciences. Washington, DC: National Academy Press, 1986
[12] Thagard P. How to make decisions: Coherence, emotion, and practical inference. In: Varieties of practical
inference. Cambridge, MA: MIT Press, 2001, pp. 355-371.
[13] The British Computer Society's Specialist Group on Expert Systems (BCS SGES)[online] URL:
http://www.chrisnaylor.co.uk/Definition.html
[14] Vroom V. H, Jago A. G. The New Leadership. Englewood Cliffs: Prentice Hall, 1988.
Kontaktní adresa:
Ing. Hana Kopáčková, Ph.D.
USII/FES Univerzita Pardubice
Studentská 84, Pardubice, 53210
e-mail: hana.kopackova@upce.cz
telefon: 466036245
ResearchGate has not been able to resolve any citations for this publication.
Conference Paper
Full-text available
Decision making processes and systems to support the same have focused for the most part on narrow disciplines, paradigms, perspectives, and pre-determined processes. Apart from these most decision processes and systems are designed to solve simple problems and are therefore unable to support complex problems that consist of interrelated decisions that span multiple domains, paradigms, and/or perspectives. To address these problems we propose conceptual decision-making and modelling processes. A flexible object-oriented decision system framework and architecture was developed and implemented to support the proposed processes. Some of the key concepts that we have been able to explore and implement are generic modelling ideas, such as data-model, model-solver, model-model, solver-visualisation, and data-visualisation independences. Specifically we have been able to explore the integration of models of different types, levels of complexity, depths of integration (aggregation, pipelining, and splicing) and orientations (satisficing as well as optimising).
Chapter
The term ‘bounded rationality’ is used to designate rational choice that takes into account the cognitive limitations of the decision-maker — limitations of both knowledge and computational capacity. Bounded rationality is a central theme in the behavioural approach to economics, which is deeply concerned with the ways in which the actual decision–making process influences the decisions that are reached.
Article
Introduction, 99. — I. Some general features of rational choice, 100.— II. The essential simplifications, 103. — III. Existence and uniqueness of solutions, 111. — IV. Further comments on dynamics, 113. — V. Conclusion, 114. — Appendix, 115.
Article
The technology for building knowledge-based systems by inductive inference from examples has been demonstrated successfully in several practical applications. This paper summarizes an approach to synthesizing decision trees that has been used in a variety of systems, and it describes one such system, ID3, in detail. Results from recent studies show ways in which the methodology can be modified to deal with information that is noisy and/or incomplete. A reported shortcoming of the basic algorithm is discussed and two means of overcoming it are compared. The paper concludes with illustrations of current research directions.